I have this Romanian 2x 10-band EQ. It needed a recap as one of the channels cut out a few months ago due to a failed coupling cap which i replaced at that time, and yesterday i had the time to take care of all of them.
I replaced all 'lytics around the transistors, signal path or not, with Rubycons. ALL of them had values all over the place. Both lower AND higher than what was printed.
If you thought you've seen bad lytics, you haven't seen Romanian ones. They were known to shift values within 1 year... and this thing is at least 20 years old.
This improved the low bass response, but i still hadn't gotten rid of the old problem - some hum would make its way to the output whenever something was connected to the input. It was only audible when the speakers were turned fairly loud, but it was enough to be annoying. The funny thing was that the hum was present whether the EQ was plugged in or not. It would go away if i would put a 100 ohm resistor between output and ground, but there went my signal as well.
So i started tracing the circuit. I have to mention that the whole thing runs off a single 40v rail, the equalization network is entirely passive (yes it's got inductors!), it only has 3 transistors per channel (presumably 1 input buffer, 1 gain stage and 1 output buffer), and a BM324 (LM324 made at Baneasa, Romania) for the clipping indicator. Obviously with a single rail the whole thing is coupled by caps.
So let's start the fun. First off, the 4.7uF output coupling cap was left unterminated. That was the source of the loud POP when the output was plugged in after the EQ was turned on - there was no path to ground and DC would build up. Here's the real kicker - we have a 3k9 resistor in series from this cap to output. WHAT??? 3.9kOhm output impedance??? No wonder it's picking up hum, and no wonder the signal went to hell with a 100 ohm to ground.
Reaching the "output buffer" transistor we find the reason for the 3k9 resistor. This transistor is biased with 2k7 resistors in both emitter and collector. Oddly enough the output signal is taken from the collector (common emitter configuration), even though a buffer stage is usually emitter follower. It doesn't matter when both resistors are the same as gain is 1 either way, but still, this is weird, and one more reason why the output impedance is high. I'll have to check absolute phase from input to output, the common emitter amp is inverting so maybe they did this to keep the signal in phase.
In the base of this transistor, we have the coupling cap from the previous stage and... that's all. No bias resistors, the base is left to bias itself... Now i see why the 40 volt power supply for something that doesn't even have to output 2 volts p-p. Well, this is gonna require some work. The transistors are TO-92 so i'll have to see if they'll take the current required to lower the output impedance enough.
On another note, the consumption of the whole thing is rated at 8 watts. Want to know why? Well... there's a big mofo dropper resistor directly from the AC line... for the power indicator bulb.
Seriously. Which bulb has been replaced by a LED a long time ago so that resistor is basically doing nothing but wasting power. Romanian engineering at its finest. 
Edit: The power resistor (4.something kohms) is actually in series with the transformer powering the whole affair. I guess they had surplus transformers... but they didn't have 230v primaries. So instead of building some new trafos they decided to add a resistor to drop enough voltage for the ones they had. Brilliant!
I replaced all 'lytics around the transistors, signal path or not, with Rubycons. ALL of them had values all over the place. Both lower AND higher than what was printed.
If you thought you've seen bad lytics, you haven't seen Romanian ones. They were known to shift values within 1 year... and this thing is at least 20 years old.This improved the low bass response, but i still hadn't gotten rid of the old problem - some hum would make its way to the output whenever something was connected to the input. It was only audible when the speakers were turned fairly loud, but it was enough to be annoying. The funny thing was that the hum was present whether the EQ was plugged in or not. It would go away if i would put a 100 ohm resistor between output and ground, but there went my signal as well.
So i started tracing the circuit. I have to mention that the whole thing runs off a single 40v rail, the equalization network is entirely passive (yes it's got inductors!), it only has 3 transistors per channel (presumably 1 input buffer, 1 gain stage and 1 output buffer), and a BM324 (LM324 made at Baneasa, Romania) for the clipping indicator. Obviously with a single rail the whole thing is coupled by caps.
So let's start the fun. First off, the 4.7uF output coupling cap was left unterminated. That was the source of the loud POP when the output was plugged in after the EQ was turned on - there was no path to ground and DC would build up. Here's the real kicker - we have a 3k9 resistor in series from this cap to output. WHAT??? 3.9kOhm output impedance??? No wonder it's picking up hum, and no wonder the signal went to hell with a 100 ohm to ground.
Reaching the "output buffer" transistor we find the reason for the 3k9 resistor. This transistor is biased with 2k7 resistors in both emitter and collector. Oddly enough the output signal is taken from the collector (common emitter configuration), even though a buffer stage is usually emitter follower. It doesn't matter when both resistors are the same as gain is 1 either way, but still, this is weird, and one more reason why the output impedance is high. I'll have to check absolute phase from input to output, the common emitter amp is inverting so maybe they did this to keep the signal in phase.
In the base of this transistor, we have the coupling cap from the previous stage and... that's all. No bias resistors, the base is left to bias itself... Now i see why the 40 volt power supply for something that doesn't even have to output 2 volts p-p. Well, this is gonna require some work. The transistors are TO-92 so i'll have to see if they'll take the current required to lower the output impedance enough.
On another note, the consumption of the whole thing is rated at 8 watts. Want to know why? Well... there's a big mofo dropper resistor directly from the AC line... for the power indicator bulb.
Seriously. Which bulb has been replaced by a LED a long time ago so that resistor is basically doing nothing but wasting power. Romanian engineering at its finest. Edit: The power resistor (4.something kohms) is actually in series with the transformer powering the whole affair. I guess they had surplus transformers... but they didn't have 230v primaries. So instead of building some new trafos they decided to add a resistor to drop enough voltage for the ones they had. Brilliant!
Comment